linux-kernel - Re: [PATCH] usb: hubs: Decrease IN-endpoint poll interval for Microchip USB491x hub

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <988f4311-a726-4a7e-b0bf-6aeec13d8f23@rowland.harvard.edu>
Date:   Thu, 23 Nov 2023 13:17:03 -0500
From:   Alan Stern <stern@...land.harvard.edu>
To:     Hardik Gajjar <hgajjar@...adit-jv.com>
Cc:     gregkh@...uxfoundation.org, corbet@....net, tj@...nel.org,
        rdunlap@...radead.org, paulmck@...nel.org,
        linux-doc@...r.kernel.org, linux-usb@...r.kernel.org,
        linux-kernel@...r.kernel.org, erosca@...adit-jv.com,
        Martin.Mueller5@...bosch.com
Subject: Re: [PATCH] usb: hubs: Decrease IN-endpoint poll interval for
 Microchip USB491x hub

On Thu, Nov 23, 2023 at 09:19:48AM +0100, Hardik Gajjar wrote:
> There is a potential delay in announcing downstream USB bus activity to
> Linux USB drivers due to the default interrupt endpoint having a poll
> interval of 256ms.
> 
> Microchip has recommended ignoring the device descriptor and reducing
> that value to 32ms, as it was too late to modify it in silicon.
> 
> This patch aims to speed up the USB enumeration process, facilitating
> the successful completion of Apple CarPlay certifications and enhancing
> user experience when utilizing USB devices through the Microchip Multihost
> Hub.
> 
> A new quirk, USB_QUIRK_REDUCE_FRAME_INTR_BINTERVAL, accelerates the
> notification process by changing the Endpoint interrupt poll interval
> from 256ms to 32ms.

But this is meant to apply only to hubs, right?  So shouldn't it be a 
HUB_QUIRK_32_MS_INTR_INTERVAL macro, used in hub.c's hub_id_table, 
rather than a general USB quirk?

> Signed-off-by: Hardik Gajjar <hgajjar@...adit-jv.com>
> ---
>  Documentation/admin-guide/kernel-parameters.txt |  4 ++++
>  drivers/usb/core/config.c                       |  8 ++++++++
>  drivers/usb/core/quirks.c                       | 11 +++++++++++
>  include/linux/usb/quirks.h                      |  5 +++++
>  4 files changed, 28 insertions(+)
> 
> diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
> index 65731b060e3f..6b0a66f0e6bf 100644
> --- a/Documentation/admin-guide/kernel-parameters.txt
> +++ b/Documentation/admin-guide/kernel-parameters.txt
> @@ -6908,6 +6908,10 @@
>  					pause after every control message);
>  				o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
>  					delay after resetting its port);
> +				p = USB_QUIRK_REDUCE_FRAME_INTR_BINTERVAL (Set
> +					bInterval to a Maximum of 9 to Reduce
> +					default Poll Rate from 256 ms to
> +					32 ms);

256 ms and 32 ms are _periods_ (or intervals), not _rates_.

bInterval=9 corresponds to 32 ms only for High Speed and SuperSpeed* 
devices.  For Low and Full Speed it corresponds to 9 ms.  Explanatory 
comments should strive not to be misleading.

Alan Stern